Crypto’s October Crash: The Hidden Story Behind $18 Billion in Liquidations

Solana Research Institute Reignites Transparency Fight After Record Meltdown

The numbers hit the crypto world like a shockwave. On the evening of October 10, 2025, roughly $18 billion in trader positions evaporated over a harrowing 14-hour stretch, with $3.21 billion wiped out in a single minute at 21:15 UTC. The scale was unprecedented, the panic visceral. But for Solana Research Institute, the collapse was not just a market catastrophe — it was evidence for a regulatory argument. In an August 14 post, the Solana-aligned research group revived a July open letter from Angus Scott to the UK Financial Conduct Authority and other regulators, holding up the crash as its central case study. The letter’s thesis was blunt: opaque, centralized trading venues crumbled under pressure while transparent, on-chain finance kept functioning. Scott’s 33-page appeal reportedly emerged from discussions between the FCA and the Solana Foundation and spanned seven regulatory domains — identity, resilience, custody, market abuse, systemic risk, prudential capital, and above all, transparency. The framing turned a market event into a referendum on infrastructure. When the selling frenzy reached its peak, traders on major centralized platforms found themselves at the mercy of lagging internal systems, delayed transfers, and pricing models that appeared to amplify the panic rather than contain it. On-chain venues, by contrast, left public footprints that let anyone follow the stress in real time. But the raw crash records, examined closely, support a far more complex conclusion. Public data allowed researchers to reconstruct a massive auto-deleveraging event on Hyperliquid, uncover deficits and oracle delays at Aave, and identify exactly how Binance’s internal collateral pricing turbocharged forced selling. The October records did not simply show a clean division between unsafe centralized venues and resilient decentralized ones; they exposed something more subtle. Transparency laid bare the mechanics of stress across the entire market structure — without turning any single category of venue into a proxy for safety.

Conflicting Numbers: The Measurement Gap Regulators Can’t Ignore

The $18 billion figure was never as solid as it appeared. Amberdata, analyzing six major exchanges, recorded $9.89 billion in liquidations across the same 14-hour window. The two totals describe different realities, and the divergence raises an uncomfortable question for policymakers: which number should guide regulatory response? Amberdata’s dataset did confirm the jaw-dropping peak — $3.21 billion liquidated in a single minute, 93.5% of it from forced selling. The same dataset showed $6.93 billion erased in just 40 minutes between 20:50 and 21:30 UTC. A separate ESMA review, meanwhile, pointed to market estimates of roughly $19 billion in automated derivatives liquidations for the entire day. Each figure is defensible; none can be reconciled with the others without far more detail. The problem is one of scope. A day-wide market estimate, a six-exchange 14-hour sample, a one-minute peak, and a venue-specific loss mechanism all answer fundamentally different questions. When they are collapsed into a single dramatic total — as Solana Research Institute’s July letter appeared to do — the market plumbing that the policy debate is meant to expose disappears from view. The letter gave no common universe of venues and no aggregation methodology, leaving the $18 billion headline disconnected from Amberdata’s $9.89 billion. What the available records actually establish is a measurement gap, not a calculation error. And that gap, arguably, is the most important discovery of the entire episode. Regulators cannot design informed policy around incomparable data. Until they can speak a shared statistical language about liquidation events, they remain unable to determine which venue failures were routine, which were systemic, and which were simply the product of design choices made in boardrooms rather than on blockchains.

Inside Binance’s Postmortem: When Internal Pricing Amplifies a Panic

Binance’s own postmortem provided a chilling case study in how a centralized platform can turn a market shock into a liquidity crisis. The exchange maintained that its spot and futures matching engines and API trading stayed fully operational throughout the chaos. But beginning at 21:18 UTC, other systems began to falter. Internal transfers and Earn redemptions slowed to a crawl. More damaging still, after 21:36 UTC the exchange’s local price feeds for collateral assets including USDe, BNSOL, and WBETH began to dislocate from genuine market value. As those internal prices de-pegged, the collateral backing trader positions appeared to evaporate, triggering forced liquidations and cascading sell orders that compounded the very panic they were meant to absorb. Binance ultimately paid out approximately $283 million across two compensation batches to users who had been liquidated because of those pricing failures. ESMA’s subsequent review did not mince words. The European regulator concluded that Binance’s use of internal collateral prices created the conditions for local de-pegs to wipe out collateral value, precipitating forced selling and a broader cascade. ESMA also noted that the turmoil produced no observable spillover into traditional financial markets — the damage, at least on that score, remained confined to crypto. But the account identifies venue design itself as a systemic amplifier, something no aggregate liquidation total can possibly communicate. What the Binance record does not provide is an event-specific auto-deleveraging number. ADL — a last-resort mechanism that trims profitable traders’ positions when ordinary liquidations and risk buffers can no longer keep a venue solvent — is distinct from the routine liquidation of a losing position. The postmortem’s own narrative sorts the night into separate components: module glitches, transfer constraints, collateral-pricing dislocations, and garden-variety forced liquidations. Each demands a different regulatory response; the headline totals capture none of them.

On-Chain Venues Felt the Stress Too — and Left a Public Record

The decentralized side of the market did not escape unbloodied. Hyperliquid, the on-chain derivatives venue, suffered a large-scale auto-deleveraging event during the same chaotic hours. A non-peer-reviewed study built on public venue data confirmed the event’s scope, establishing that ADL — not just ordinary liquidation — occurred at significant volume on one of the sector’s most prominent transparent platforms. The finding matters because ADL is fundamentally different: it doesn’t merely close a losing position after collateral falls below a threshold; it actively reduces the positions of profitable traders as a matter of last resort, to keep the venue solvent when risk buffers have been exhausted. Regulators need comparable records to separate ADL from an outage, an oracle delay, or a venue-local pricing failure. Hyperliquid’s disclosure made such analysis possible — but it also demonstrated that transparency is not a shield. Aave’s experience told a similar story with different machinery. A report from Chaos Labs documented price-update delays of up to five blocks in some Aave markets during the crash. The protocol absorbed deficits that, under stressed conditions, would have been far more damaging; Chaos Labs estimated that liquidation fees and SVR revenue ultimately left Aave roughly $1.5 million net positive. Yet the episode exposed the vulnerability of oracle-dependent lending in real time. Public records allowed outsiders to reconstruct segments of Hyperliquid’s loss allocation and Aave’s lending stress — a genuine victory for observability. But observability did not erase the losses; it simply made them legible. The on-chain venues failed differently, recovered differently, and disclosed differently. Comparing them with Binance and other centralized platforms requires preserving those distinctions. The moment they are flattened into a single narrative of centralized failure and decentralized success, the policy debate loses the very detail it needs most.

FCA’s New Rules Are a Step Forward, but the Blind Spots Linger

The regulatory landscape is already shifting in response to the tension the crash exposed. The FCA’s final cryptoasset framework, published in June 2026, requires qualifying UK cryptoasset trading platforms and principal dealers to publish post-trade information as close to real time as possible and no later than one minute after execution. Larger UK platform operators face additional pre-trade transparency obligations. The rules mark a meaningful advance for crypto market transparency. Yet their reach into decentralized finance remains deliberately limited: the framework applies to DeFi only where a clear controlling person is carrying out regulated cryptoasset activity. Genuinely decentralized operations sit outside the perimeter for now, with a separate consultation on DeFi guidance still pending. That cautious carve-out reflects the recurring difficulty of applying venue-based regulation to infrastructure without a headquarters. More consequential, perhaps, is what the framework does not require. It imposes no standardized, cross-venue reporting of liquidation volumes, ADL usage, or backstop losses. Faster trade data sharpens the view of execution — the what, when, and where of individual transactions — but the October records show how operational delays, pricing failures, and loss-allocation mechanisms can remain stubbornly difficult to compare after a common shock. Twenty-three minutes of glitching modules at Binance, five-block oracle delays at Aave, a single-minute liquidation peak, a $283 million compensation claim: each is a data point in a different format, aimed at a different question. Without a shared standard for venue-level event reporting, regulators will continue assembling the story of a crisis from fragments. The FCA’s framework, welcome as it is, does not yet bridge that gap. Solana Research Institute’s 33-page letter addresses the full arc of that problem, positioning the October 10 crash not as an isolated event but as a case study in what still cannot be seen.

The Real Lesson of the Crash: Transparency Is Visibility, Not Safety

The strongest element of Solana Research Institute’s argument is precisely the one that resists sloganizing. The October crash proved that public records can make venue failures measurable, including failures on transparent platforms. Hyperliquid’s ADL, Aave’s oracle delays, Binance’s collateral-price dislocations — all of it was reconstructable because some party left a trail. The observability gap is real, and the institute is right to push regulators toward closing it. What the data cannot support is the tidy conclusion that centralized venues are dangerous and on-chain venues are safe. The records show transparent infrastructure experiencing its own stress, and centralized infrastructure revealing critical information about its own failures. Transparency is not proof of safety; it is proof of visibility. And visibility, essential as it is, is only where regulation begins. The task for the FCA, ESMA, and their global counterparts is not to crown a winning category of market infrastructure — October 10, 2025, offers no such clean verdict. The real work is to construct comparable event disclosures that allow regulators to distinguish routine solvency controls from venue-specific operational or pricing breakdowns, whether the venue runs on a blockchain or in a server farm. Until liquidation volumes, ADL triggers, oracle delays, and backstop losses are reported in standardized formats across every platform, regulators will continue solving an equation with missing variables. The debate over whether $18 billion or $9.89 billion is the true total matters less than the components that remain unseen. That, more than any single number, is the lesson of the crash. And it is the reason the conversation Solana Research Institute started — flawed numbers, contested conclusions, and all — is one the industry cannot afford to abandon.